Medical image processing method, device, processing equipment, and storage medium

Abstract

The embodiments of the present application relate to a medical image processing method, apparatus, processing equipment, and storage medium, wherein the method includes: acquiring a brain magnetic resonance image within an acquisition time period; extracting time series signals of a plurality of voxels contained in the brain magnetic resonance image within the acquisition time period; performing multimodal decomposition on the time series signals corresponding to the plurality of voxels, and respectively determining the modal signal sets corresponding to the plurality of voxels; respectively calculating the similarity between the modal signal sets of the plurality of voxels, and dividing the plurality of target voxels in the plurality of voxels whose similarity meets preset requirements into the same brain network. The medical image processing method provided in the embodiments of the present application can perform multimodal decomposition on the time series signals of a plurality of voxels in parallel to determine the modal signal sets corresponding to the time series signals, thereby reducing calculation time and improving the analysis efficiency of brain magnetic resonance images.

Description

Technical Field

[0001] The present application relates to the field of medical technology, and in particular to a medical image processing method, apparatus, processing device, and storage medium. Background Art

[0002] Functional magnetic resonance imaging (fMRI) is a non-invasive neuroimaging method that uses magnetic resonance imaging to measure changes in blood dynamics caused by neuronal activity. In recent years, the focus of neuroscience research has shifted significantly to studying the brain in the "resting state," focusing on intrinsic brain activity in the absence of any sensory or cognitive stimulation. Analysis of resting-state brain functional connectivity has revealed distinct resting-state networks that describe specific functions and distinct spatial topologies.

[0003] In related technologies, the Independent Component Correlation Algorithm (ICA) method is generally used to extract functional brain networks. Specifically, the fMRI data can be rearranged into a matrix of spatial and temporal dimensions, and the matrix can be decomposed to obtain a series of different brain networks and extracted noise components. However, since resting-state fMIR is a set of 4-dimensional data combined in time and space, when there are many time points collected or the resolution is high, especially at ultra-high fields, the matrix size will increase exponentially. The longer the ICA decomposition matrix takes, the higher the computing power required of the computer, which may cause memory overflow.

[0004] Therefore, related technologies urgently need a high-accuracy medical image processing method to extract brain networks. Summary of the Invention

[0005] Based on this, in response to the above technical problems, the present application provides a medical image processing method, apparatus, processing equipment, and storage medium, which can solve the problem of inaccurate extraction of functional brain networks in related technologies.

[0006] In a first aspect, an embodiment of the present application provides a medical image processing method, the method comprising:

[0007] obtaining brain magnetic resonance images within an acquisition time period;

[0008] extracting time series signals of a plurality of voxels contained in the brain magnetic resonance image within the acquisition time period;

[0009] Performing multimodal decomposition on the time series signals corresponding to the plurality of voxels, and respectively determining modal signal sets corresponding to the plurality of voxels;

[0010] The similarities between the modal signal sets of the plurality of voxels are respectively calculated, and a plurality of target voxels among the plurality of voxels whose similarities meet preset requirements are divided into the same brain network.

[0011] The medical image processing method provided in the embodiment of the present application can extract the time series signals of the multiple voxels contained in the brain magnetic resonance image within the acquisition time period. Then, the time series signals corresponding to the multiple voxels can be subjected to multimodal decomposition to determine the modal signal sets corresponding to the multiple voxels. Finally, by calculating the similarity between the modal signal sets of the multiple voxels, the whole-brain functional network in the resting state can be constructed. Compared with the existing technology, the speed of extracting brain networks is faster and more accurate. In addition, since the time series signals corresponding to the multiple voxels can be decomposed in parallel to determine the corresponding modal signal sets during the multimodal decomposition process, the calculation time can be reduced and the analysis efficiency of brain magnetic resonance images can be improved.

[0012] Optionally, in one embodiment of the present application, performing multimodal decomposition on the time series signals corresponding to the plurality of voxels to respectively determine the modal signal sets corresponding to the plurality of voxels includes:

[0013] The time series signals of the multiple voxels are subjected to multimodal decomposition, and at least one signal whose signal frequency obtained by decomposition is less than a preset frequency threshold is constructed into a modal signal set corresponding to each voxel, wherein the preset frequency threshold is determined based on an empirical value of the brain signal frequency.

[0014] Optionally, in one embodiment of the present application, performing multimodal decomposition on the time series signals corresponding to the plurality of voxels to respectively determine the modal signal sets corresponding to the plurality of voxels includes:

[0015] Adding a white noise signal to the time series signal to construct a time series signal to be decomposed;

[0016] The time series signal to be decomposed is subjected to multimodal decomposition using an empirical mode decomposition method to obtain multi-order intrinsic mode signals, and the multi-order intrinsic mode signals are constructed as a set of modal signals corresponding to the time series signal.

[0017] Optionally, in one embodiment of the present application, respectively calculating the similarities between the modal signal sets of the plurality of voxels, and dividing the plurality of target voxels whose similarities meet preset requirements into the same brain network, includes:

[0018] Determining signals having the same modality in a set of modality signals corresponding to different voxels;

[0019] The similarity between the modality signal sets of the plurality of voxels is determined according to the similarity between the signals having the same modality.

[0020] Optionally, in one embodiment of the present application, determining the similarity between the modality signal sets of the plurality of voxels based on the similarity between the signals having the same modality includes:

[0021] Performing Fourier transform on the signals with the same mode to determine frequency spectrum information corresponding to the signals with the same mode;

[0022] The similarity between the modal signal sets of the plurality of voxels is determined based on the spectral coherence between the spectral information corresponding to the signals having the same modality.

[0023] Optionally, in one embodiment of the present application, after respectively calculating the similarities between the modal signal sets of the plurality of voxels and dividing the plurality of target voxels whose similarities meet preset requirements into the same brain network, the method further includes:

[0024] Functionally connect the multiple target voxels that are divided into the same brain network to construct a brain functional network.

[0025] In a second aspect, an embodiment of the present application further provides a medical image processing device, the device comprising:

[0026] A magnetic resonance image acquisition module is used to acquire brain magnetic resonance images within an acquisition time period;

[0027] A timing signal extraction module is used to extract timing signals of a plurality of voxels contained in the brain magnetic resonance image within the acquisition time period;

[0028] a decomposition module, configured to perform multimodal decomposition on the time series signals corresponding to the plurality of voxels, and respectively determine modal signal sets corresponding to the plurality of voxels;

[0029] The similarity calculation module is used to respectively calculate the similarities between the modal signal sets of the multiple voxels, and divide the multiple target voxels whose similarities meet preset requirements among the multiple voxels into the same brain network.

[0030] In a third aspect, an embodiment of the present application further provides a processing device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the methods described in the above embodiments when executing the computer program.

[0031] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium having computer program instructions stored thereon, which implement the steps of the methods described in the above embodiments when the computer program instructions are executed by a processor.

[0032] In a fifth aspect, an embodiment of the present application further provides a chip comprising at least one processor, wherein the processor is configured to run computer program instructions stored in a memory to execute the steps of the methods described in the above embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments or descriptions of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0034] Figure 1 This is a schematic diagram of an application scenario provided by an embodiment of the present application;

[0035] Figure 2 A flowchart of a medical image processing method provided in one embodiment of the present application;

[0036] Figure 3 is a schematic diagram of extracting a timing signal corresponding to a voxel provided by an embodiment of the present application;

[0037] Figure 4 is a schematic diagram of an empirical mode decomposition provided in an embodiment of the present application;

[0038] Figure 5 This is a schematic diagram of a functional connection provided by an embodiment of the present application;

[0039] Figure 6 1 is a schematic diagram of the module structure of the medical image processing device 103 provided in an embodiment of the present application;

[0040] Figure 7 It is a schematic diagram of the module structure of the processing device 700 provided in an embodiment of the present application. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is described and illustrated below in conjunction with the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. Based on the embodiments provided in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application. In addition, it can also be understood that although the efforts made in this development process may be complex and lengthy, for ordinary technicians in the field related to the contents disclosed in the present application, some changes such as design, manufacturing or production based on the technical contents disclosed in the present application are only conventional technical means and should not be understood as the contents disclosed in the present application being insufficient.

[0042] References to "embodiments" in this application mean that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it refer to independent or alternative embodiments that are mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described in this application may be combined with other embodiments unless there is a conflict.

[0043] Unless otherwise defined, the technical or scientific terms used in this application should have the usual meanings understood by persons of ordinary skill in the technical field to which this application belongs. The words "one", "a", "the" and the like used in this application do not indicate a limit on quantity and may indicate the singular or plural. The terms "include", "comprise", "have" and any variations thereof used in this application are intended to cover non-exclusive inclusions; for example, a process, method, system, product or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units that are not listed, or may also include other steps or units that are inherent to these processes, methods, products or devices. The words "connect", "connected", "coupled" and the like used in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The term "multiple" used in this application means greater than or equal to two. The terms "first", "second", "third" and the like used in this application are merely used to distinguish similar objects and do not represent a specific ordering of objects.

[0044] In addition, numerous specific details are provided in the following detailed description to better illustrate the present application. Those skilled in the art will appreciate that the present application can be practiced without certain specific details. In some instances, devices, means, components, and circuits well known to those skilled in the art are not described in detail in order to highlight the main purpose of the present application.

[0045] Task state and resting state are two states of the brain. The former mainly refers to the state of the brain when it is performing specific tasks such as memory, recognition and movement. The resting state refers to the state when the brain is not performing specific cognitive tasks, remains quiet, relaxed and awake. In practical applications, magnetic resonance imaging (MRI) equipment can be used to collect brain scan data based on functional magnetic resonance imaging (fMRI) technology. Functional magnetic resonance imaging is a non-invasive technology that indirectly detects neuronal activity through hemodynamic signals. Resting state fMRI (rs-fMRI) data refers to fMRI image data collected when the brain is not performing specific cognitive tasks, remains quiet, relaxed and awake.

[0046] Currently, resting-state fMRI data can be analyzed, for example, by calculating the temporal correlation (synchrony) of activity across different brain regions to construct a resting-state whole-brain functional network. There are two main methods for resting-state fMRI analysis: seed point functional connectivity analysis and independent component analysis (ICA). This method performs blind source decomposition on 4-dimensional resting-state fMRI data based on the assumption of spatial independence, resulting in the identification of various functional networks. However, when the number of acquisition time points is large or the resolution is high, the processing efficiency of these two methods decreases, and the burden on the processing equipment increases.

[0047] Based on actual technical requirements similar to those described above, the present application provides a medical image processing method that can extract the time series signals of multiple voxels contained in the brain magnetic resonance image within the acquisition time period. Then, the time series signals corresponding to the multiple voxels can be subjected to multimodal decomposition to determine the modal signal sets corresponding to the multiple voxels. Finally, the similarities between the modal signal sets of the multiple voxels are calculated respectively to construct the whole-brain functional network in the resting state. Compared with the prior art, the method of determining the similarity between multiple voxels by the similarity between multiple modal signal sets has faster calculation speed and higher calculation accuracy. In addition, since the time series signals corresponding to the multiple voxels can be decomposed in parallel to determine the corresponding modal signal sets during the multimodal decomposition process, the calculation time can be reduced and the analysis efficiency of resting-state fMRI data can be improved.

[0048] See Figure 1 , Figure 1: This is a schematic diagram of an application scenario provided by an embodiment of the present application. The application scenario may include a magnetic resonance imaging device 101 and a medical image processing device 103, wherein the magnetic resonance imaging device 101 and the medical image processing device 103 can communicate to send the acquired brain magnetic resonance image to the medical image processing device 103, and the medical image processing device 103 completes the division of the brain network. The medical image processing device 103 can use the medical image processing method provided in any of the following embodiments of the present application to process the brain magnetic resonance image acquired by the magnetic resonance imaging device 101. Specifically, after acquiring the brain magnetic resonance image, the medical image processing device 103 can extract the time series signals of the multiple voxels contained in the brain magnetic resonance image within the acquisition time period, and perform multimodal decomposition on the time series signals corresponding to the multiple voxels to determine the modal signal sets corresponding to the multiple voxels. Finally, based on the similarity between the modal signal sets of the multiple voxels, the multiple target voxels whose similarity meets the preset requirements (for example, greater than the preset similarity threshold) in the multiple voxels can be divided into the same brain network. The medical image processing device 103 may be an electronic device with data processing and data transmission and reception capabilities. The electronic device may be a physical device or a cluster of physical devices, such as a server or a server cluster. Of course, the electronic device may also be a virtualized cloud device, such as at least one cloud computing device in a cloud computing cluster. This application does not impose any restrictions on the form of the electronic device.

[0049] The medical image processing method described in this application is described in detail below with reference to the accompanying drawings. Figure 2 : It is a flow chart of an embodiment of the medical image processing method provided by the present application. Although the present application provides the method operation steps as shown in the following embodiments or drawings, more or fewer operation steps may be included in the method based on routine or no creative labor. In the steps where there is no necessary causal relationship logically, the execution order of these steps is not limited to the execution order provided in the embodiment of the present application. In the actual medical image processing process or when the method is executed, the method can be executed in the order of the method shown in the embodiment or drawings or in parallel (for example, in a parallel processor or multi-threaded processing environment).

[0050] Specifically, an embodiment of the medical image processing method provided by this application is as follows: Figure 2 As shown, the method may include:

[0051] S201: Acquire a brain magnetic resonance image within an acquisition time period.

[0052] In an embodiment of the present application, the magnetic resonance imaging device 101 can be used to scan the brain of a target subject during an acquisition period to determine a brain magnetic resonance image of the target subject. The target subject can be a human or animal, for example. The acquisition period can be a time range set by the user based on acquisition requirements. For example, the target subject's brain can be scanned between 8:00 AM and 8:10 AM. The user can be a doctor, nurse, clinical pathologist, medical imaging specialist, radiologist, sonographer, etc. Specifically, the type of brain magnetic resonance image acquired varies depending on the target subject's brain state. Specifically, when the target subject's brain is in a task state, the acquired brain magnetic resonance image is a task state magnetic resonance image; when the target subject's brain is in a resting state, the acquired brain magnetic resonance image is a resting state magnetic resonance image. In one embodiment of the present application, to determine the target subject's whole-brain functional connectivity in the resting state, the brain magnetic resonance image described in each of the following embodiments can be a resting state magnetic resonance image. Specifically, the brain magnetic resonance image can include information on the time-varying hemodynamic signals of multiple voxels at different locations in the brain. Based on this, the brain magnetic resonance image is a four-dimensional image, comprising three-dimensional spatial dimensions and a time dimension. The three-dimensional spatial dimensions can be represented by x, y, and z, and the time dimension can be represented by T. In one embodiment of the present application, the storage format of the brain magnetic resonance image can include DICOM format, NIFTI format, and the like.

[0053] During the acquisition process, differences in the subject's weight, height, or acquisition location, as well as physiological activities such as head movement, breathing, and heartbeat, may introduce interference factors that may affect subsequent analysis. Based on this, in one embodiment of the present application, the brain MRI image can be preprocessed to remove noise, increase the signal-to-noise ratio, and so on. Specifically, the preprocessing of the brain MRI image can include slice timing, head motion correction, spatial normalization, spatial smoothing, and so on. For example, during the acquisition process, the subject's head will inevitably move, and spontaneous physiological activities such as breathing and heartbeat may also cause head movement. Therefore, head motion correction is necessary to reduce the impact of head motion on the data. For example, it can be set to exclude data collected at the corresponding moment when the head moves in a certain direction by more than 0.5 mm or rotates by more than 1 degree. In one embodiment of the present application, the spatial normalization can be performed by deforming and adjusting the entire brain to standardize it to a standard brain. For example, the brain magnetic resonance image may be spatially transformed using an EPI template to achieve spatial normalization.

[0054] S203: Extracting time series signals of a plurality of voxels included in the brain magnetic resonance image within the acquisition time period.

[0055] In the embodiment of the present application, since the brain magnetic resonance image may include information on the changes in the hemodynamic signals of multiple voxels at different locations in the brain over time, specifically, the hemodynamic signal may be a blood oxygenation level dependent (BOLD) signal. Therefore, the information on the changes in the BOLD signals of multiple voxels contained in the brain magnetic resonance image over time may be extracted separately to form a time series signal of the multiple voxels within the acquisition time period. In other words, the time series signal may include a BOLD signal that changes with the acquisition time. For example, in one example, as Figure 3 As shown, the BOLD signal y(t) of a voxel in the brain magnetic resonance image that changes with time can be determined, wherein the horizontal axis can be used to represent the acquisition time t, and the vertical axis can be used to represent the BOLD signal.

[0056] S205: Perform multimodal decomposition on the time series signals corresponding to the plurality of voxels, and respectively determine modal signal sets corresponding to the plurality of voxels.

[0057] In practical applications, the time series signals corresponding to multiple voxels belong to nonlinear and non-stationary signal sequences. Therefore, when the similarity between multiple time series signals is subsequently calculated, the task is relatively complex and the accuracy is not high. Based on this, in an embodiment of the present application, the time series signals corresponding to the multiple voxels can be first subjected to multimodal decomposition to determine the multiple modal signals corresponding to the time series signals. Since the time series signal can be composed of several modal signals such as intrinsic mode functions, the multimodal decomposition is to use a modal decomposition method to decompose the time series signal into multiple modal signals to determine the multiple modal signals contained in the time series signal. Then, a modal signal set corresponding to the voxel can be constructed based on the multiple modal signals. In this way, the similarity between multiple voxels can be quickly and accurately determined by calculating the similarity between different modal signal sets. Among them, the modal signal can be a local characteristic signal of different time scales containing the time series signal. In one embodiment of the present application, the time series signals corresponding to the multiple voxels can be subjected to multimodal decomposition using a time-frequency signal decomposition method. That is, the multimodal decomposition may include decomposing the time series signal into multiple modal signals. The time-frequency signal decomposition method may include empirical mode decomposition (EMD), ensemble empirical mode decomposition (EEMD), complementary ensemble empirical mode decomposition (CEEMD), complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN), etc.

[0058] It should be noted that when performing multimodal decomposition on time series signals corresponding to multiple voxels, the multimodal decomposition of the multiple time series signals can be performed in parallel or serially. In one embodiment of the present application, in order to improve computational efficiency, the multimodal decomposition of the multiple time series signals can be performed in parallel.

[0059] In actual applications, the signal frequency of the brain network is usually in the range of 0.01kHz-0.1kHz. In general, signals with a frequency higher than 0.1kHz can be considered as physiological noise such as the breathing and heartbeat of the target object. Therefore, in order to improve the accuracy of calculating the similarity between the modal signal sets corresponding to multiple voxels to improve the accuracy of dividing the brain network, the modal signals that do not meet the requirements can be eliminated from the multiple modal signals obtained by decomposition after multimodal decomposition. Specifically, in one embodiment of the present application, the time series signals corresponding to the multiple voxels are subjected to multimodal decomposition, and the modal signal sets corresponding to the multiple voxels are respectively determined, including:

[0060] S301: Perform multimodal decomposition on the time series signals of the multiple voxels, and construct at least one signal whose signal frequency obtained by decomposition is less than a preset frequency threshold into a modal signal set corresponding to each voxel, wherein the preset frequency threshold is determined based on an empirical value of the brain signal frequency.

[0061] In an embodiment of the present application, after performing multimodal decomposition on the time series signals of the multiple voxels, at least one modal signal whose frequency of the decomposed modal signal is greater than a preset frequency threshold can be eliminated. The preset frequency threshold can be set by the user based on the empirical value of the brain signal frequency, and the brain signal may include the BOLD signal of the brain. For example, many studies have shown that the frequency of the BOLD signal of the brain fluctuates in the range of 0.01-0.1kHz. Therefore, the preset frequency threshold can be set to 0.1kHz. After eliminating the modal signals greater than the preset frequency threshold from the multiple modal signals, the remaining modal information less than the preset frequency threshold can be constructed to generate a modal signal set corresponding to each voxel.

[0062] It should be noted that when performing multimodal decomposition on the time series signals of multiple voxels, there may be a voxel where the signal frequencies of all the signals decomposed from the time series signal are greater than the preset frequency threshold. In other words, the modal signals contained in the time series signal corresponding to this voxel are all noise signals. To simplify the complexity of subsequent calculations and improve computational efficiency, this voxel can be eliminated and the subsequent similarity calculation of the modal signal set is not performed, thereby improving the efficiency and accuracy of brain network segmentation.

[0063] In the above embodiment, after performing multimodal decomposition on the time series signals of the multiple voxels, modal signals with a frequency greater than a preset threshold can be removed. This can remove background noise, allowing for more accurate calculation of the similarity between multiple modal signal sets, thereby improving the accuracy of brain network segmentation.

[0064] In another embodiment of the present application, in order to improve the efficiency and accuracy of background noise removal, a method of set empirical mode decomposition can be used to perform multimodal decomposition on the time series signals of the multiple voxels. Specifically, performing multimodal decomposition on the time series signals corresponding to the multiple voxels and determining the modal signal sets corresponding to the multiple voxels respectively include:

[0065] S401: Adding a white noise signal to the time series signal to generate a time series signal to be decomposed;

[0066] S403: Perform multimodal decomposition on the time series signal to be decomposed using an empirical mode decomposition method to obtain multi-order intrinsic mode signals, and construct the multi-order intrinsic mode signals into a modal signal set corresponding to the time series signal.

[0067] In an embodiment of the present application, in the process of performing multimodal decomposition on the time series signals of the plurality of voxels using the method of collective empirical mode decomposition, a white noise signal can be first superimposed on the time series signal to construct a time series signal to be decomposed. The white noise signal can be a Gaussian signal whose frequency is uniformly distributed, and the amplitude of the white noise signal can be set in advance. After determining the time series signal to be decomposed, the empirical mode decomposition method can be used to perform multimodal decomposition on the time series signal to be decomposed. In one embodiment of the present application, Figure 4 As shown, the multiple maximum points contained in the time series signal to be decomposed s(t) can be first determined as shown in Figure (a), and the upper envelope of the time series signal to be decomposed can be formed by fitting using a difference function such as a cubic spline difference function. Afterwards, the lower envelope of the time series signal to be decomposed can be determined based on the multiple minimum points contained in the time series signal to be decomposed in the same way. Then, the mean curve of the upper envelope and the lower envelope can be obtained as shown in Figure (b). Finally, the original data sequence s(t) can be subtracted from the mean curve to determine a new time series signal h1(t) as shown in Figure (c). In the case that there are no negative local maximums and positive local minimums in the new time series signal, it can be determined that the new time series signal is the first intrinsic mode signal. If there are negative local maximums and positive local minimums in the new time series signal, the above processing is continued for the new time series signal. It can be understood that since the time series signal corresponding to the voxel contains the change information of the BOLD signal corresponding to the voxel, that is, the BOLD signal y(t), the time series signal to be decomposed can also be the BOLD signal y(t). Correspondingly, the empirical mode decomposition method described in the above embodiment can be used to perform multimodal decomposition on the time series signal y(t) to be decomposed.

[0068] In an embodiment of the present application, after performing multimodal decomposition on the time series signal to be decomposed using the empirical mode decomposition method to obtain the first intrinsic mode signal, the above-mentioned steps of superimposing the white noise signal to continue the empirical mode decomposition can be repeated, and a white noise signal with the same amplitude can be added each time to obtain eigenmode signals of different orders. Finally, the eigenmode signals of different orders can be averaged to determine the final decomposition result. It can be seen that in the process of performing multimodal decomposition on the time series signals of the multiple voxels using the collective empirical mode decomposition method, the time series signal can be superimposed on the white noise signal with a consistent distribution throughout the entire time-frequency space, so that the modal signals of different time scales will be automatically distributed to the appropriate reference scale. In addition, after multiple averaging, the noise will cancel each other out. In this way, the noise signal can be screened out in advance, and there is no need to perform the next step of similarity calculation, thereby improving the efficiency and accuracy of dividing the brain network.

[0069] Through the above embodiment, the time series signals of the multiple voxels can be subjected to multimodal decomposition using the method of collective empirical mode decomposition to determine multi-order intrinsic mode signals. This can construct a set of modal signals corresponding to each voxel, thereby helping to improve the accuracy of multimodal decomposition and more effectively obtain signal features. In addition, during this process, noise signals in the time series signals can be automatically removed, thereby improving the efficiency and accuracy of noise signal removal.

[0070] S207: Calculate the similarities between the modal signal sets of the plurality of voxels respectively, and divide the plurality of target voxels whose similarities meet preset requirements into the same brain network.

[0071] In an embodiment of the present application, after determining the modal signal sets corresponding to multiple voxels, the similarity between the multiple modal signal sets can be determined. In one embodiment of the present application, since the modal signal set contains multiple modal signals, when calculating the similarity between the modal signal sets, the similarity between the modal signal sets can be determined by the similarity between the multiple modal signals in the modal signal set and the multiple modal signals in another modal signal set. For example, in one example, modal signal set A contains modal signals m(t) and n(t), and modal signal set B contains modal signals p(t) and q(t). Then the similarity between the modal signal set A and the modal signal set B can be determined based on the similarity between the modal signal m(t) and the modal signal p(t) and the similarity between the modal signal n(t) and the modal signal q(t). Specifically, the algorithm for determining the similarity can include a distance calculation method, a cosine similarity calculation method, a kernel function calculation method, and a Pearson correlation coefficient. The distance calculation method may include a Minkowski distance calculation method, a Euclidean distance calculation method, a Mahalanobis distance, and the like. In one embodiment of the present application, after determining the similarity between the multiple modal signal sets, the target voxels corresponding to the multiple modal signals whose similarities meet the preset requirements can be divided into the same brain network. Specifically, in one embodiment of the present application, the situation where the similarity meets the preset requirements may include that the similarity is the maximum value among the multiple candidate similarities, or may include that the similarity is greater than a preset similarity threshold. The preset similarity threshold can be set by the user according to actual application requirements, for example, it can be set to 95%, 90%, and so on. In some other embodiments of the present application, multiple candidate similarities can also be selected, and then the average value of the multiple candidate similarities is calculated, and the average value is determined as the preset similarity threshold. The present application does not impose any restrictions on the setting of the preset similarity threshold.

[0072] The medical image processing method provided in the embodiment of the present application can extract the time series signals of the multiple voxels contained in the brain magnetic resonance image within the acquisition time period. Then, the time series signals corresponding to the multiple voxels can be subjected to multimodal decomposition to determine the modal signal sets corresponding to the multiple voxels. Finally, by calculating the similarity between the modal signal sets of the multiple voxels, the whole-brain functional network in the resting state can be constructed. Compared with the existing technology, the speed of extracting brain networks is faster and more accurate. In addition, since the time series signals corresponding to the multiple voxels can be decomposed in parallel to determine the corresponding modal signal sets during the multimodal decomposition process, the calculation time can be reduced and the analysis efficiency of brain magnetic resonance images can be improved.

[0073] In one embodiment of the present application, in order to improve the accuracy of calculating the similarity between multiple modal signal sets, the similarity between the multiple modal signal sets can be determined based on the similarity between signals of the same modality in the modal signal sets corresponding to different voxels. Specifically, the similarity between the modal signal sets of the multiple voxels is calculated separately, and multiple target voxels in the multiple voxels whose similarity meets preset requirements are divided into the same brain network, including:

[0074] S501: Determine signals having the same modality in a set of modality signals corresponding to different voxels;

[0075] S503: Determine the similarity between the modality signal sets of the plurality of voxels according to the similarity between the signals having the same modality.

[0076] In an embodiment of the present application, when performing multimodal decomposition on the time series signals corresponding to the multiple voxels, multiple modal signals corresponding to the time series signals can be determined, and the frequencies and frequency change trends corresponding to different modal signals are different. Therefore, the multiple modal signals corresponding to the multiple modal signal sets can be extracted separately. Then, based on the frequencies and frequency change trends corresponding to different modal signals, the modal signals with the same modality in the modal signals corresponding to different voxels can be determined. Thereafter, the similarity between the modal signal sets of the multiple voxels can be determined based on the similarity between the signals with the same modality. For example, in one example, modal signal set A includes modal signals m(t) and n(t), and modal signal set B includes modal signals p(t) and q(t). Among them, the modal signal m(t) and the modal signal q(t) are the same modality, and the modal signal n(t) and the modal signal p(t) are the same modality. Then the similarity between the modal signal set A and the modal signal set B can be determined based on the similarity between the modal signal m(t) and the modal signal q(t) and the similarity between the modal signal n(t) and the modal signal p(t). Of course, in another example of the present application, the modal signals m(t) and n(t) can be spliced to obtain a first signal sequence, and the modal signals q(t) and p(t) can be spliced to obtain a second signal sequence. Thereafter, the similarity between the modal signal set A and the modal signal set B can be determined based on the similarity between the first signal sequence and the second signal sequence. Specifically, in one embodiment of the present application, determining the similarity between the modal signal sets of the multiple voxels based on the similarity between the signals having the same modality includes:

[0077] S601: Performing Fourier transform on the signals with the same mode to determine frequency spectrum signals corresponding to the signals with the same mode respectively;

[0078] S603: Determine the similarity between the modal signal sets of the plurality of voxels according to the spectral coherence between the spectral signals corresponding to the signals having the same modality.

[0079] In an embodiment of the present application, the signals having the same modality may be subjected to a Fourier transform to determine the spectral signals corresponding to the signals having the same modality, that is, the signals having the same modality may be converted from the time domain to the frequency domain. For example, in one example, the modal signal x(t) may be converted into a spectral signal F(ω) using the Fourier transform formula. The Fourier transform formula is shown below:

[0080]

[0081] In one embodiment of the present application, after determining the spectral signals corresponding to the signals having the same modality, the similarity between the modal signal sets of the multiple voxels can be determined based on the spectral coherence between the various spectral signals. In one embodiment of the present application, the similarity between the various spectral signals can be calculated using the Pearson correlation coefficient calculation method. The Pearson correlation coefficient calculation method can be used to determine the degree of linear correlation between the various spectral signals. Specifically, the degree of linear correlation can be calculated using the following formula:

[0082]

[0083] Among them, the It can be the average value of the first spectrum signal, It can be the average value of the second spectrum signal. i can be used to represent the amplitude of each component contained in the first spectrum signal, the x iIt can be used to represent the amplitude of each component contained in the second spectrum signal. In one embodiment of the present application, the similarity between the first spectrum signal and the second spectrum signal can be determined based on the size of the correlation coefficient r. For example, if the correlation coefficient r is in the range of 0.8-1.0, the first spectrum signal and the second spectrum signal are extremely similar; if the correlation coefficient r is in the range of 0.0-0.2, the first spectrum signal and the second spectrum signal are extremely similar. In one embodiment of the present application, when the correlation coefficient is greater than 0.8, it can be determined that the spectral coherence between the first spectrum signal and the second spectrum signal meets the preset requirements, and the first voxel corresponding to the first spectrum signal and the second voxel corresponding to the second spectrum signal are divided into the same brain network. Of course, in other embodiments of the present application, other similarity calculation methods such as cosine similarity calculation method, mutual information calculation method, etc. can also be used to determine the spectral coherence between the signals with the same modality, and this application is not limited here.

[0084] Through the above embodiment, the signals with the same modality can be converted into corresponding spectral signals through Fourier transform. In this way, the similarity between the modal signals can be converted into corresponding spectral coherence, so that the similarity between the modal signal sets can be more easily calculated later.

[0085] In practical applications, the brain is a network composed of different brain regions, each with its own tasks and functions. By analyzing the functional connectivity characteristics between different brain regions, the brain's language, memory and other functions can be examined, located and explored. Therefore, in one embodiment of the present application, after respectively calculating the similarity between the modal signal sets of the multiple voxels and dividing the multiple target voxels whose similarity meets the preset requirements into the same brain network, it also includes:

[0086] S701: Functionally connect the multiple target voxels that are divided into the same brain network to construct a brain functional network.

[0087] In an embodiment of the present application, functional connectivity (FC) can be performed on multiple target voxels divided into the same brain network. The functional connectivity can establish a connection relationship between multiple target voxels with the help of linear time correlation. After functionally connecting multiple voxels in each brain network, the brain functional network can be constructed. Among them, the brain functional network may include multiple networks, for example, it may include a salience network, an auditory network, a basal ganglia network, a high-level visual network, a visual-spatial network, a default mode network, a language network, an executive network, a precuneus network, a primary visual network, a sensorimotor network, and the like. For example, in one example, as Figure 5 As shown, each target voxel 501 in the region 500 can be functionally connected to construct a brain functional network.

[0088] The above describes in detail the medical image processing method provided by this application. Figure 6 , describing the medical image processing apparatus 103 provided in the present application, the medical image processing apparatus 103 may include:

[0089] The magnetic resonance image acquisition module 1031 is used to acquire a brain magnetic resonance image within a collection time period;

[0090] A timing signal extraction module 1033 is configured to extract timing signals of a plurality of voxels contained in the brain magnetic resonance image within the acquisition time period.

[0091] a decomposition module 1035 , configured to perform multimodal decomposition on the time series signals corresponding to the plurality of voxels, and respectively determine modal signal sets corresponding to the plurality of voxels;

[0092] The similarity calculation module 1037 is configured to calculate the similarity between the modal signal sets of the plurality of voxels, and to classify the plurality of target voxels whose similarity satisfies a preset requirement into the same brain network.

[0093] Optionally, in one embodiment of the present application, performing multimodal decomposition on the time series signals corresponding to the plurality of voxels to respectively determine the modal signal sets corresponding to the plurality of voxels includes:

[0094] The time series signals of the multiple voxels are subjected to multimodal decomposition, and at least one signal whose signal frequency obtained by decomposition is less than a preset frequency threshold is constructed into a modal signal set corresponding to each voxel, wherein the preset frequency threshold is determined based on an empirical value of the brain signal frequency.

[0095] Optionally, in one embodiment of the present application, performing multimodal decomposition on the time series signals corresponding to the plurality of voxels to respectively determine the modal signal sets corresponding to the plurality of voxels includes:

[0096] Adding a white noise signal to the time series signal to construct a time series signal to be decomposed;

[0097] The time series signal to be decomposed is subjected to multimodal decomposition using an empirical mode decomposition method to obtain multi-order intrinsic mode signals, and the multi-order intrinsic mode signals are constructed as a set of modal signals corresponding to the time series signal.

[0098] Optionally, in one embodiment of the present application, respectively calculating the similarities between the modal signal sets of the plurality of voxels, and dividing the plurality of target voxels whose similarities meet preset requirements into the same brain network, includes:

[0099] Determining signals having the same modality in a set of modality signals corresponding to different voxels;

[0100] The similarity between the modality signal sets of the plurality of voxels is determined according to the similarity between the signals having the same modality.

[0101] Optionally, in one embodiment of the present application, determining the similarity between the modality signal sets of the plurality of voxels based on the similarity between the signals having the same modality includes:

[0102] Performing Fourier transform on the signals with the same mode to determine frequency spectrum information corresponding to the signals with the same mode;

[0103] The similarity between the modal signal sets of the plurality of voxels is determined based on the spectral coherence between the spectral information corresponding to the signals having the same modality.

[0104] Optionally, in one embodiment of the present application, after respectively calculating the similarities between the modal signal sets of the plurality of voxels and dividing the plurality of target voxels whose similarities meet preset requirements into the same brain network, the method further includes:

[0105] Functionally connect the multiple target voxels that are divided into the same brain network to construct a brain functional network.

[0106] According to the embodiments of the present application, the medical image processing device 103 may correspond to executing the method described in the embodiments of the present application, and the above-mentioned and other operations and / or functions of each module in the medical image processing device 103 are respectively for implementing the corresponding processes of the methods provided in the above-mentioned embodiments. For the sake of brevity, they will not be repeated here.

[0107] It should also be noted that the embodiments described above are merely illustrative, wherein the modules described as separate components may or may not be physically separate, and the components displayed as modules may or may not be physical modules, that is, they may be located in one place, or they may be distributed across multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment. In addition, in the drawings of the device embodiments provided in this application, the connection relationship between the modules indicates that there is a communication connection between them, which can be specifically implemented as one or more communication buses or signal lines.

[0108] On the other hand, the present application further provides a processing device, including a memory and a processor, wherein the memory stores computer program instructions, and the processor is configured to run the computer program instructions to execute the medical image processing method described in each of the above embodiments.

[0109] The processing device may be a physical device or a cluster of physical devices, or a virtualized cloud device, such as at least one cloud computing device in a cloud computing cluster. For ease of understanding, this application uses the processing device as an independent physical device to illustrate the structure of the processing device.

[0110] like Figure 7 As shown, the processing device 700 includes: a processor and a memory for storing processor computer program instructions; wherein the processor is configured to implement the above-mentioned device when executing the computer program instructions. The processing device 700 includes a memory 701, a processor 703, a bus 705 and a communication interface 707. The memory 701, the processor 703 and the communication interface 707 communicate with each other through the bus 705. The bus 705 can be a peripheral component interconnect standard (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 7 The bus is represented by only one thick line, but it does not mean that there is only one bus or one type of bus. The communication interface 707 is used for external communication.

[0111] The processor 703 may be a central processing unit (CPU). The memory 701 may include a volatile memory, such as a random access memory (RAM). The memory 701 may also include a non-volatile memory, such as a read-only memory (ROM), a flash memory, a HDD, or an SSD.

[0112] Those skilled in the art will understand that Figure 7 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0113] On the other hand, the present application further provides a chip, comprising at least one processor, wherein the processor is configured to run computer program instructions stored in a memory to execute the steps of the methods described in the above embodiments.

[0114] On the other hand, the present application further provides a computer-readable storage medium having computer program instructions stored thereon, which implement the steps of the methods described in the above embodiments when the computer program instructions are executed by a processor.

[0115] A computer-readable storage medium can be a tangible device that can hold and store instructions used by an instruction execution device. A computer-readable storage medium can be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination thereof. More specific examples (non-exhaustive list) of computer-readable storage media include: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disc (DVD), a memory stick, a floppy disk, a mechanical encoding device, a punch card or a raised structure in a groove on which instructions are stored, and any suitable combination thereof.

[0116] The computer program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer program instructions from the network and forwards the computer program instructions to be stored in the computer-readable storage medium in the respective computing / processing device.

[0117] The computer program instructions for performing the operations of the present application may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages such as Smalltalk, C++, and conventional procedural programming languages such as "C" or similar programming languages. The computer program instructions may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, by utilizing the status information of computer program instructions to personalize an electronic circuit, such as a programmable logic circuit, a field-programmable gate array (FPGA), or a programmable logic array (PLA), the electronic circuit can execute the computer program instructions, thereby implementing various aspects of the present application.

[0118] Various aspects of the present application are described herein with reference to flowcharts and / or block diagrams of methods and apparatus according to embodiments of the present application. It should be understood that each block of the flowcharts and / or block diagrams, and combinations of blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions.

[0119] These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine, so that when these instructions are executed by the processor of the computer or other programmable data processing device, a device is generated to implement the functions / actions specified in one or more blocks in the flowchart and / or block diagram. These computer program instructions can also be stored in a computer-readable storage medium, where these instructions cause the computer, programmable data processing device, and / or other equipment to operate in a specific manner. Thus, the computer-readable medium storing the instructions comprises an article of manufacture, which includes instructions for implementing various aspects of the functions / actions specified in one or more blocks in the flowchart and / or block diagram.

[0120] Computer program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operational steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to implement the functions / actions specified in one or more blocks in the flowchart and / or block diagram.

[0121] The flow charts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the devices, systems and methods according to multiple embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a part of a module, program segment or instruction, and the part of the module, program segment or instruction comprises one or more executable instructions for realizing the logical function of the specification. In some alternative implementations, the functions marked in the box can also occur in a sequence different from that marked in the accompanying drawings. For example, two continuous boxes can actually be executed substantially in parallel, and they can sometimes also be executed in the opposite order, depending on the functions involved.

[0122] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A medical image processing method, characterized in that: The method comprises: obtaining brain magnetic resonance images within an acquisition time period; extracting time series signals of a plurality of voxels contained in the brain magnetic resonance image within the acquisition time period; Performing multimodal decomposition on the time series signals corresponding to the plurality of voxels, and respectively determining modal signal sets corresponding to the plurality of voxels; The similarities between the modal signal sets of the plurality of voxels are respectively calculated, and a plurality of target voxels among the plurality of voxels whose similarities meet preset requirements are divided into the same brain network.

2. The method according to claim 1, characterized in that The performing multimodal decomposition on the time series signals corresponding to the plurality of voxels to respectively determine the modal signal sets corresponding to the plurality of voxels includes: The time series signals of the multiple voxels are subjected to multimodal decomposition, and at least one signal whose signal frequency obtained by decomposition is less than a preset frequency threshold is constructed into a modal signal set corresponding to each voxel, wherein the preset frequency threshold is determined based on an empirical value of the brain signal frequency.

3. The method according to claim 1 or 2, characterized in that The performing multimodal decomposition on the time series signals corresponding to the plurality of voxels to respectively determine the modal signal sets corresponding to the plurality of voxels includes: Adding a white noise signal to the time series signal to construct a time series signal to be decomposed; The time series signal to be decomposed is subjected to multimodal decomposition using an empirical mode decomposition method to obtain multi-order intrinsic mode signals, and the multi-order intrinsic mode signals are constructed as a set of modal signals corresponding to the time series signal.

4. The method according to claim 1, wherein The respectively calculating similarities between the modal signal sets of the plurality of voxels and dividing the plurality of target voxels whose similarities meet preset requirements into the same brain network includes: Determining signals having the same modality in a set of modality signals corresponding to different voxels; The similarity between the modality signal sets of the plurality of voxels is determined according to the similarity between the signals having the same modality.

5. The method according to claim 4, characterized in that Determining the similarity between the modal signal sets of the plurality of voxels according to the similarity between the signals having the same modality includes: Performing Fourier transform on the signals with the same mode to determine frequency spectrum information corresponding to the signals with the same mode; The similarity between the modal signal sets of the plurality of voxels is determined based on the spectral coherence between the spectral information corresponding to the signals having the same modality.

6. The method according to claim 1, characterized in that After respectively calculating the similarities between the modal signal sets of the plurality of voxels and dividing the plurality of target voxels whose similarities meet preset requirements into the same brain network, the method further includes: Functionally connect the multiple target voxels that are divided into the same brain network to construct a brain functional network.

7. A medical image processing device, characterized in that: The device comprises: A magnetic resonance image acquisition module is used to acquire brain magnetic resonance images within an acquisition time period; A timing signal extraction module is used to extract timing signals of a plurality of voxels contained in the brain magnetic resonance image within the acquisition time period; a decomposition module, configured to perform multimodal decomposition on the time series signals corresponding to the plurality of voxels, and respectively determine modal signal sets corresponding to the plurality of voxels; The similarity calculation module is used to respectively calculate the similarities between the modal signal sets of the multiple voxels, and divide the multiple target voxels whose similarities meet preset requirements among the multiple voxels into the same brain network.

8. A processing device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A chip, characterized in that: The system comprises at least one processor configured to execute computer program instructions stored in a memory to perform the steps of the method according to any one of claims 1 to 6.

Images